Course details

Research Topics in ML and DL

Research Topics in ML and DL

Expected Duration
Lesson Objectives
Course Number
Expertise Level


This course explores research being done in machine learning and deep learning. Topics covered in its 12 videos include neural networks and deep neural networks. First, learners examine how to prevent neural networks from overfitting. You will explore research on multilabel learning algorithms, multilabel classification, and multiple-output classifications, which are variants of the standard classification problem. Then examine deep learning algorithms, and how the depth of neural networks is key to performance since deeper neural networks have been shown to be more adept at automatic feature extraction. Next, learners will examine research involved in transferable features in deep neural networks, and research associated with large-scale video classification. You will review research related to common objects in context, and generative adversarial networks. You will learn about research on facial alignment, and ensemble of regression trees, and about deep features for scene recognition. Finally, you will look at research proposing Extreme Learning Machine (ELM), which is used for regression and multiclass classification.

Expected Duration (hours)

Lesson Objectives

Research Topics in ML and DL

  • Course Overview
  • understand the efforts being undertaken to reduce overfitting using the dropout technique
  • understand leading edge multi-label learning algorithms
  • understand the proposed learning framework for deep residual learning that improves training of networks that are significantly deeper than traditional neural networks
  • understand how initializing a network with transferred features may boost generalization performance
  • understand how convolutional neural networks may be utilized as a powerful class of models for image recognition
  • understand the dataset that advances state-of-the-art object recognition by considering the context within the question of scene understanding
  • understand the proposed framework for estimating generative models via an adversarial process that successfully estimates the probability that a sample came from training data rather than a generative model
  • understand how optimal nearest neighbor algorithms perform compared to traditional nearest neighbor algorithms
  • understand how an ensemble of regression trees may successfully estimate facial landmark positions while delivering real-time performance and high quality predictions
  • understand how a proposed new scene-centric database is successfully used for learning deep features for scene recognition
  • recognize how ELM tends to produce better scalability, generalization performance, and faster learning than traditional support vector machine
  • understand the trending research topics in ML and DL
  • Course Number:

    Expertise Level